Assessing, perceiving and insuring credit risk
This thesis is concerned with the assessment, perception and insurance of credit risk. The thesis aims to make contributions both within these areas, and at specific points of interface between them. No attempt is made to develop a single unifying thesis. Rather, a series of partial models are developed, both theoretical and empirical, that develop and connect particular facets of financial economics. The first model demonstrates how movements in market risk produce movements in lender risk-assessment effort. It is demonstrated that deleterious movements in market-wide risk can actually produce a fall in assessment effort. The capricious nature of risk assessment causes changes in the lender's perception of the weights placed on determinants. This has important implications for borrowers' attempts to minimize risk premiums. Time-variability of signal-weights is tested using structural break tests on ordinary least squares and fixed effects panel models. Results suggest a fluid relationship between risk and determinants. Central to empirical investigation is the measurement of perceived risk. A critique of potential measures rejects the use of interest rate spreads - the most commonly used measure - on the basis that they do not take into account the possibility of credit rationing. A model is then constructed to reproduce the standard explanation of credit rationing - Adverse Selection induced Credit Rationing Equilibrium (ASCRE). This model is then extended to include classificatory risk assessment. Assessment is found to reduce the scope for ASCRE, and to cause favourable selection. Credit insurance is then included, and it is found that insurance cover makes risk assessment less of an imperative to lenders, and reduces the utility losses from raising interest rates. The parallel implication is that credit insurance weakens ASCRE, to the extent that full insurance with flat-rate premiums removes the possibility of ASCRE altogether. If the terms of insurance are made contingent on the terms of the loan, a new form of credit rationing emerges: Contingent Insurance induced Credit Rationing Equilibrium (CICRE). CICRE is separate, but not mutually exclusive, to ASCRE. A theoretical model of the demand for loan insurance is developed, and empirically estimated, in the context of the UK mortgage market. Inter alia, the model examines the role of auto-perception of risk determining credit insurance demand. Results reveal the take-up of credit insurance to be relatively insensitive to the borrower's perception ofhis/her own risk.